Can we measure the effect of retrospectives?

This question popped into my mind during Josef's Futurespectives session [link needed] on the Monday.

Some of the outcomes of retrospectives are easily quantifiable - you can track defect rates or the change in code coverage. However, the most important (for me at least) aspects aren't so readily measured. So, I wondered rather than measuring those, can we measure the effect a retrospective has? Is it worth the time? I didn't go into the session with anything thought through; if anything I expected people to say “no you can't” and walk off - if they came at all.

I'll add the photos when Flickr lets me back in

Thanks to all who came.

Hosted by John McFadyen